[mikeh]

kgr, on Jan 10 2009, 04:34 AM, said:

I did setup a deal profile in Jack with bidding and with given cards for opening leader.
I dealt a deal and Jack did bid as requested in the deal profile, so contract was 3Dx.
Analyse position for lead:
♠J: -17.5
♥4: -44.0
♠9: -48.6
♣2: -53.4
♦5: -191.6
- I'm not sure how this works and what Jack uses for possible deals for this bidding (maybe does does not correspond to biddigns we would do).
- It seems like the exxpected score for this deal is negative for us.

if your constraints resulted in, on average, a minus score, I suggest that something is very wrong with either or both your constraints or the program that did the analysis

[nige1]

kgr, on Jan 10 2009, 04:34 AM, said:

I did setup a deal profile in Jack with bidding and with given cards for opening leader.
I dealt a deal and Jack did bid as requested in the deal profile, so contract was 3Dx.
Analyse position for lead:
♠J: -17.5
♥4: -44.0
♠9: -48.6
♣2: -53.4
♦5: -191.6
- I'm not sure how this works and what Jack uses for possible deals for this bidding (maybe does does not correspond to biddigns we would do).
- It seems like the exxpected score for this deal is negative for us.

mikeh, on Jan 10 2009, 02:18 PM, said:

if your constraints resulted in, on average, a minus score, I suggest that something is very wrong with either or both your constraints or the program that did the analysis

Perhaps, because the expected score from 3♦X is less than the score of a game for us. Unfortunagtely, opponent's pre-empt deprives us of the bidding space to find the right game, so we do the best we can in the circumstances.

[dburn]

I would lead a spade. After all, if I blow a trick in spades I may get it back by scoring the eight of diamonds on a promotion.

If you told me I could not lead a spade, I would lead a heart. If you told me I could not lead either major, I would bid 3NT.

[y66]

Anything but a diamond.

Per 500 deal simulation, expected results in descending order are

H2 555
H4 555
H7 555

SJ 553
S9 551

C2 551
C6 551
C9 550
CQ 548

D2 424
D5 424
D8 369
DJ 163

In this sample, a spade lead produced 1100 slightly more often than other leads (23 times out of 500 vs 19 times for heart leads and 17 times for club leads). But it also produced more 100s and 300s at the other end of the histogram. And in one case, a spade lead let them make 3DX when North had AK4 AKQT 64 AKJ4 and East had Q8632 8 AKQT973 --.

Simulation specs:

North: balanced, 22-24
East: 7 diamonds (can't have 8)
South: as given
West: the rest

Hands generated by Jack. Leads analyzed using PM Cronje's and Bo Haglund's Double Dummy software,

[FrancesHinden]

I think some of these simulations are a bit off.
East only had 6 diamonds, and the 3D overcall was found at both tables.
(Edit: may I should have mentioned that the overcaller is a junior, and team-mate at the other table is a 64-year-old junior-at-heart)

The good news is that, on the actual hand, any lead except for the obviously insane jack of diamonds leads to +800, so in some sense this is not an interesting problem.

I posted it because my partner led a club, which I was slightly surprised by. I would have led a spade.

At the other table they justified my teammate's 3D overcall by having no idea how to bid the strong hand over intervention, and ended up in 6NT-3.

[paulg]

FrancesHinden, on Jan 12 2009, 10:28 AM, said:

The good news is that, on the actual hand, any lead except for the obviously insane jack of diamonds leads to +800, so in some sense this is not an interesting problem.

I posted it because my partner led a club, which I was slightly surprised by. I would have led a spade.

Any problem that gets Fred and Mike disagreeing as violently as they ever will is clearly a good one.

All we really needed at the end was JLOL to promote a heart.